Design And Implementation Of Malicious Detection System For Android Application

With the incredible development of mobile Internet technology, an enormous amount of amazing APPs are provided by the open operating platform which supporting open API. However, its openness character also makes a lot of malicious applications appeared in the application market. Android operating system is one of the most popular system for smart mobile devices, while its openness is highly appreciated, but it also face the problem that the appearance of widely spread of malicious applications. There are two reasons for the widely spread of malicious applications:on one hand, it’s easy for anyone to develop an APP since the Android is not hard even for a new developer; on the other hand, there are too many application download source so that not all of them could have a nice regulatory mechanism. These make the malicious application leaking personal information, sending message to some SMS service or visit some website without the user’s permission. All of these kinds of operations damaged user’s interests. So how to avoid the application doing bad things to the users of smart mobile devices by detecting the malicious actions becomes a very big problem in mobile Internet platform field.For the security requirements of Android operating system, this article provide a static analysis method based on DE compilation and characteristic code detection from the view of security analysis. An online Android APP malicious detection system has been designed and implemented in this paper. The system will do the DE compilation operation with the uploaded APK file according to the selected DE compilation program, and make the byte code file into CFG. Then use the analyze tool to judge whether the APP is safe or not. Finally, the system extracts related characteristic code as output result. The system could change different DE compilation program and analyze tool to do the whole analysis process, and product a result log to users.According to the final test results, our system could detect most kind of malicious behavior. The analysis process will be finished in a very short time when we first execute the analysis on the APP, and it will be much faster if the analysis is the repeated execution, so we could say the system is efficient.